基于多智能体的城市群政策协调建模与仿真

Modeling and Simulation of Multi-Cities' Policy Coordination Based on Mas

以城市群多政府互动与政策偏好演化为例,通过概念模型、数学模型和计算机模型的完整建模过程构建一个多智能体仿真模型,基于动力系统理论构建微观决策主体的偏好演化机制,基于复杂网络模型模拟宏观社会网络的拓扑演化规则,并对仿真模型和模拟结果进行效度和信度的检验。模拟实验探讨了全局交互周期比例、局部交互连接概率、行政/激励调控措施及其交叉作用对城市群政策协调演化的动态影响,并结合大样本模拟数据的统计分析,为促进城市群协调合作和区域一体化进程提供决策依据和政策参考,是多智能体建模与仿真(ABMS)在公共管理和政策领域的前沿拓展,提供了新的研究视角和方法论体系,也是政府组织模拟实验研究(行政学科计算化与实验化)的一次新尝试。

To analyze the Multi-Cities＇ Government Interaction and Policy Preference Evolution, a multi-a- gent simulation system is built through conceptual model, mathematic model and computer model. The preference evolutionary mechanism of micro decision-making entities is constructed based on dynamic sys- tem, and the topology evolving rule of macro social networks is built based on complex networks, and the reliability and validity test for simulation model and results is given. The simulation experiments investi- gate the effect of global interaction proportion, local interlinking probability, administration/stimulation measures and their interaction on the evolving process of the multi-cities＂ policy coordination. The statistics analysis of large sample of simulation data is integrated, to provide policy advices to advance the collaboration and integration of urban agglomeration. This research is a frontier expansion of the multi-agent sys- tem in the field of public management and policy, endowing the discipline with new research perspective and methodological system, and also an innovative try of government organizations simulation experiment research.